Martijn Ernst

RUNX1-Familial Platelet Disorder (RUNX1-FPD) is a rare inherited disease characterized by a bleeding propensity due to platelet defects, and a predisposition to develop hematological malignancies, mostly acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS). Clinical characteristics and pathogenesis of hematologic malignancy caused by RUNX1-FPD remain poorly understood. In this thesis, we investigated clinical and biological characteristics of RUNX1-FPD, with particular interest for AML occurring in its context.

Firstly, chapter 1 forms an introduction to key aspects of biology and medicine that are relevant for the studies described in this thesis. As such, this chapter provides an introduction to the biology of normal and malignant hematopoiesis, including the role of the bone marrow microenvironment (niche) in which adult hematopoiesis is mainly localized. Furthermore, clinical aspects of myeloid malignancies are introduced, with a focus on AML and MDS. Additionally, the concept of germline predisposition to hematologic malignancies is discussed, including multiple examples of diseases to illustrate the clinical heterogeneity of this disease entity and its clinical consequences. Lastly, biological and clinical aspects of RUNX1-FPD, the main focus of this thesis, are introduced.

In chapter 2, we investigated optimal clinical management of AML with germline predisposition, including RUNX1-FPD, by reviewing recent guidelines and expert opinion. With respect to identification of RUNX1-FPD in AML patients, literature showed agreement that performing germline diagnostics (including pre- and post-test counseling) should be considered in cases with personal or familial history of (hematologic) malignancy and/or platelet defects. Additionally, identifying RUNX1 mutations with a variant allele frequency (VAF) exceeding 40% on (somatic) sequencing panels in the context of standard AML diagnostic work-up, warrants germline diagnostics.

Regarding treatment of RUNX1-FPD AML, broad consensus exists to perform allogeneic stem cell transplantation from a non-RUNX1-FPD donor (either unrelated, or related after genetic exclusion of the germline RUNX1 variant). General consensus for follow-up includes yearly peripheral blood (PB) counts, with abnormalities warranting bone marrow (BM) puncture.

In chapter 3, we sought to determine the frequency of RUNX1 germline variants among RUNX1-mutant AML cases. We compared paired in-house HOVON sequencing data from diagnostic and complete remission (CR) samples in a trial cohort of RUNX1-mutant AML. Germline variants are identified by their persistence in CR, typically with a VAF that approximates 50%. Here, we determined that 3/37 (8.1%) RUNX1-mutated AML patients harbored a germline variant.

In chapter 4, we investigated clinical characteristics of RUNX1-FPD, in particular of AML in its context. Through international collaboration, we established a European-wide retrospective cohort of 161 patients. In this cohort, 60 out of 134 eligible patients (44.8%) had developed a hematological malignancy, with a relatively young median age (35 and 40.5 years old for AML and MDS, respectively). In AML cases (n=40), we found that co-occurring somatic events regularly involve RUNX1 (either as a mutation or a +21 or del21q chromosomal abnormality), with other most frequently mutated genes including FLT3(-ITD), SRSF2 and BCOR, and other most frequently affected chromosomes including chromosomes 7 (-7/del7q) and 8 (+8). Interrogating outcome data revealed that 20 out of 25 (80%) eligible patients achieved CR after induction therapy, with a 5-year overall survival (OS) of 50.4%. Survival in this cohort of AML with a germline RUNX1 variant was statistically superior when compared to a historical cohort of adult RUNX1-mutant, and a trend towards superior outcome remained after correction for age and treatment differences.

In chapter 5, we investigated how steady-state hematopoiesis is affected by RUNX1 deficiency in non-hematopoietic ancillary cells, one of the characteristics that distinguish RUNX1-FPD patients that develop AML from patients sporadically developing AML with a somatic RUNX1 mutation. We exploited a mouse model combining germline RUNX1 hypomorphic and loss-of-function alleles to generate an array of functional gene dosage levels. We showed that this model captures the key characteristic of reduced platelets, as well as increased myeloid potential and decreased lymphoid potential in mice with germline RUNX1 deficiency. An apparent gene dosage effect was observed, as phenotype severity increased with decreasing RUNX1 gene dosage levels. Additionally, a competitive disadvantage in reconstituting lymphoid lineages was observed for RUNX1-deficient BM, in line with relative defects in lymphopoiesis. These abnormalities were not observed upon targeted RUNX1 deletion in BM mesenchymal stromal cells (MSCs), or after transplantation of normal BM in RUNX1-deficient mice. Taken together, our data indicated hematopoietic cell autonomous disruption of steady-state hematopoiesis by RUNX1 deficiency. As such, findings do not support the model of niche-induced leukemogenesis in RUNX1-FPD, although a facilitating role of niche components in RUNX1-FPD pathogenesis was not formally excluded.

Finally, Chapter 6 provides an overall discussion of the findings in this thesis in light of pre-existent scientific literature and potential future directions.